Dr Eike Rinke co-authors paper on 'Expert-Informed Topic Models for Document Set Discovery'
The paper, co-authored by Dr Eike Rinke, has been published in the Journal of Communication Methods and Measures.
Dr Rinke’s paper proposes a new way of combining the depth of human expert knowledge with the power of machine learning for text analysis for social sciences. The research paper will help offer insight across social science disciplines that have undertaken a 'computational turn' in research methods.
When talking about the paper, Dr Rinke explained that, “The outputs from paper can be applied to ‘text-as-data’ studies that use ‘big data’ from less familiar countries and (sub-)cultures.
“The paper makes the wider point about the need to combine computational methods in the social sciences with deep social knowledge of humans.
“Specifically, the paper offers a new approach to identifying public debates of research interest and documents belonging to these debates in unfamiliar contexts. We have benchmarked, tested and validated the method by finding articles about the public role of religion from a wide set of Australian, Swiss, and Turkish blog posts and provided researchers with a complete workflow to help them apply the method to their own work”.